Real-time exciton dynamics with time-dependent density-functional theory

TL;DR

This paper presents a real-time TDDFT approach for modeling exciton dynamics in materials, aligning with linear response results and enabling femtosecond-scale simulations following short pulses.

Contribution

It introduces a method to simulate exciton dynamics in real-time TDDFT using an xc vector potential derived from long-range corrected kernels, bridging linear response and nonlinear regimes.

Findings

Consistent with frequency-dependent linear response

Access to femtosecond exciton dynamics

Potential extension into nonlinear regime

Abstract

Linear-response time-dependent density-functional theory (TDDFT) can describe excitonic features in the optical spectra of insulators and semiconductors, using exchange-correlation (xc) kernels behaving as $-1/k^{2}$ to leading order. We show how excitons can be modeled in real-time TDDFT, using an xc vector potential constructed from approximate, long-range corrected xc kernels. We demonstrate for various materials that this real-time approach is consistent with frequency-dependent linear response, gives access to femtosecond exciton dynamics following short-pulse excitations, and can be extended with some caution into the nonlinear regime.